Resatisfy Insights
Turn thousands of experiences into answers.
A dashboard tells you the number went down. Insights tells you which location, which day, which hour, which theme, and what to change — and then lets you ask a follow-up in plain language.
- One Customer Experience Score per location, composed from satisfaction, wait, recovery and engagement
- AI observations that name the window, the driver and the highest-impact fix
- Ask Resatisfy anything — every answer is computed from your own rows
- A capacity simulator that turns a satisfaction dip into a staffing decision you can defend
Experience score
0.0+4.3%
Customer visits
0+3.1%
Avg satisfaction
0.0/5+2.9%
Average wait
0 min−8.2%
AI insights
Satisfaction drops 8% on Saturdays between 11 AM and 2 PM
71% of negative feedback in that window references wait times. Adding service capacity is the highest-impact change available.
Fairfax is trailing the group by 24 CX points
Average wait is 34 minutes against a group average of 24, with 44 recovery cases still open.
Capabilities
What you actually get.
- Customer Experience Score
- 45% satisfaction, 25% wait performance against your own target, 20% recovery rate, 10% feedback engagement, with a direct drag from negative feedback share.
- Location comparison
- Every site on one score with a health status, so a fourteen-point gap is visible in one glance instead of five reports.
- AI insights
- "Satisfaction at Fairfax drops 8% on Saturdays between 11 AM and 2 PM. 71% of negative feedback in that window references wait times."
- Ask Resatisfy
- Ask why satisfaction fell, which location needs attention, or who your highest-rated employees are. Answers are computed from aggregates, then written up.
- The wait tail
- Median, 90th and 95th percentile, not just the average — because the tail is what generates the complaints.
- Topic intelligence
- Positive themes, negative themes and complaint categories, with share of voice and trend, extracted from every verbatim.
- Pressure map
- A weekday-by-hour grid of demand against the capacity that was actually working, scored on how far each hour ran past that location's target wait.
- Capacity simulator
- Add or remove people in a window and read the projected wait, tail, satisfaction, breach count and CX score back — fitted to that location's own history.
How it works
End to end, in order.
4 stages · one record per visit
- 01
Capture
Every check-in, wait, service, message, rating and case is written to the customer experience graph as it happens.
- 02
Aggregate
Metrics are computed on demand against the operational tables with an equal-length comparison window for every trend.
- 03
Observe
Deterministic rules surface the weakest location, the weakest window, the dominant complaint and the standout employee.
- 04
Decide
Take the worst window into the simulator, price the fix in people and minutes, and take a number to the conversation instead of a hunch.
Capacity simulator
Price the fix before you argue about it.
Everyone already knows Saturday lunchtime is bad. The argument is about what one more person on the desk is actually worth. The simulator finds the hours where demand outruns the desk, fits a queueing model to that location's own history, and reads back what a staffing change does to the wait, the tail, satisfaction and the location's CX score.
- Demand comes from your check-ins and your recorded service durations, not from assumptions
- An M/M/c queueing model with three parameters fitted per location, including how much of a headcount actually reaches the queue
- The before figures are measured, never modelled — the model is only used to move off them
- Satisfaction is regressed against minutes of wait over target on your own rated visits
- Every step is shown in the product under How this is calculated, so the number can be checked
Saturday · 11 AM – 2 PM · Fairfax BMW
+1 advisorWait against target, by hour — as measured
Average wait
34m—
90th percentile wait
58m—
Projected satisfaction
3.87/5—
Past the 45m breach line
214—
Location CX score
65.2—
Fitted on this location's own history · 1,842 visits across 26 weekday-hour cells. A projection from historical rows, not a guarantee — the product shows every step of the calculation, needs at least thirty rated visits in the window before it will read satisfaction from it, and says so when a window is too thin to trust.
Compare every location on one number.
A group operator does not need five dashboards. They need to know which site is dragging the group down this month and why — and to see it before the monthly review, not after it.
Tysons
Healthy0
CX score
Group average 77
Alexandria
Watch0
CX score
Group average 77
Fairfax
Warning0
CX score
Group average 77
Questions
The things buyers actually ask.
- Where do the numbers come from?
- The operational tables. There is no separate analytics pipeline to drift out of sync, so a number on the dashboard always reconciles with the rows behind it.
- Can Ask Resatisfy make things up?
- No. It classifies the question, runs the matching aggregate queries and answers from those values. A model only rewrites the phrasing, never the figures.
- How much history does the capacity simulator need?
- Enough to fit on. It needs weekday-hour cells with at least eight measured visits each, and it says on screen when a location has too few — in that case the before figures are still real but the projection falls back to generic defaults, and the product tells you not to quote it.
- Does it predict satisfaction from a handful of ratings?
- It will not. Below thirty rated visits inside the window it says so, keeps the real average of the responses it has, and borrows the wait-to-satisfaction coefficient from the wider organisation fit rather than inventing one.
- Is the projection a guarantee?
- No, and the product says so in the same panel. It is a projection from historical rows that assumes arrival volume and service mix hold. Where an hour is saturated enough that the queue never settles, the figure is bounded and labelled as “worse than this” rather than read literally.
- Can we export it?
- Yes. Every customer, visit, feedback response and case is exportable, and the REST API exposes the same aggregates the dashboards use.
- How far back does it go?
- As far back as your data. Comparison windows are always equal-length, so a 30-day view is measured against the previous 30 days.